Synopsis: process swi1:net gives 100% CPU usage.

[Sergey Pronin]

Synopsis: process swi1:net gives 100% CPU usage.

Not depending on the conditions (no heavy load, not a lot of traffic passing
through, not a lot of ng nodes) server stops to work properly.
1) swi1:net gives me 100% CPU usage.
2) server is not responding to icmp echo requests
3) ssh of course not working
4) mpd has an "ngsock" state at the top
5) tasq of the em0 card using 0% of the cpu.
6) rebooting the server helps.

What do I have:
3 simillar servers.
INTEL S3200SH with Q8200.
NIC: 82571EB using default em driver (6.9.6)
OS: Freebsd 7.1-RELEASE-p2
Soft: mpd 4.4.1 using netgraph modules, ipfw + dummynet, pf (nat only).
PPPoE.
Using only em0 card.
About 200 vlans.
2000 ng nodes created.
About 300-400 simultaneous PPPoE sessions. When error occurs there may be
about 100 PPPoE sessions.
I'm using tablearg, so there are not a lot of ipfw rules. (About 20)


sysctl.conf:

net.inet.ip.intr_queue_maxlen=1000
net.inet.tcp.blackhole=2
net.inet.udp.blackhole=1

net.inet.ip.dummynet.hash_size=1024
net.inet.ip.dummynet.io_fast=1
net.inet.ip.fw.one_pass=1

net.inet.ip.fastforwarding=1
kern.ipc.maxsockbuf=1048576

net.isr.direct=0
#net.inet.ip.portrange.randomized=0
net.inet.tcp.syncookies=1

net.inet.ip.portrange.first=1024
net.inet.ip.portrange.last=65535

dev.em.0.rx_processing_limit=1024



loader.conf:

autoboot_delay="2"
kern.ipc.maxpipekva=10000000
vm.kmem_size=512M
net.inet.tcp.tcbhashsize=8096
kern.maxusers=512
kern.ipc.nmbclusters=16384
kern.ipc.maxsockets=16384
net.graph.maxalloc=2048
net.graph.maxdgram=1048576
net.graph.recvspace=1048576
hw.em.rxd="512"
hw.em.txd="512"

And I have the 4th server with the same hardware and software configuration
but with FreeBSD 7.1-RELEASE-p1. Everything works fine.

Any ideas?
Thank you.